How I "twisted" behavioral factors

Welcome, Gabriele! Today I'd like to tell you about a very intriguing thing, as well as behavioral factors. Or rather, about their artificial improvement. More precisely, about the shortcomings of the cheaters who do this.

Quite a bit about behavioral factors

Istochnik

Why am I writing this? The fact that I am the owner of two commercial sites and, of course, I am interested that they are as high as possible in search results. And at some point, after reading various articles, I decided to experiment. Immediately say, that the result was somewhat disappointing and I hope that this article will serve as a warning.

Start

This:
the

the
• 2 commercial website;
the
• Theme: construction;
the
• Region promotion: Moscow;
the
• Age of the sites: one 5 years old, the second 3 years;
the
• Traffic up to the start of the experiment: site a — ~350 uses./day, site B is ~200./day;
the
• Start / end of experiment: January-April 2014;
the
• Links are not purchased from November 2013;

For more “purity” I have not made any changes on the website, in addition to posting articles once every one to two weeks. Notice that I'm going to compare the position before and after, and traffic, as for me personally, this is a more important metric.

Select service cheating

Over this moment for a long time did not have to think — decided to use the most actively advertised and the new service at the time. The link will not give, but let me just say that he has an interesting name and logo. :)

promotion Strategy

After sitting on the forums and after reading articles from multiple sources I was pleasantly the following decision:

the

the
• To add A website to 1500 crossings per month
the
• On site B add 900 crossings per month

That is, I added approx 15% of the current monthly traffic. I think that most will understand that to cheat is the number of transitions 1-3 queries somewhat dubious, I took 80-90 NCH and midrange queries for each site.

Disadvantages cheat: depth and time display

During the project settings in the service, I was immediately jarred by the lack of depth setting viewing and time on the website.

I believe that each website on a specific topic has its own depth and time of viewing the website. If my sites average depth is 4 pages, but time is 4-5 minutes, the shaft of the visits with a depth of 2-3 pages and a time for 2-3 minutes once out of the General statistics. In the end it turned out.

Disadvantages cheat: characteristics of visitors

In Yandex.The metric is very wonderful and informative sections on “Gender and age” and “around the world”.

So, if you see ““around the world” has always prevailed in Russia, the Center (and slightly Northwest), after a month of promotion, I saw the Ural, South, Siberia, Volga region. Also, active participants of the CIS: Ukraine, Belarus, Kazakhstan and a lot of other things.

The section “Gender and age” is typical of my sites picture: the main audience is men 25 years of age. After a month of promotion was to grow the female audience, and the audience is under the age of 18.

Well, it's not business failure? Of course, the support service has been actively convince me of this after I sent them the ticket with the comments. The essence of perebiranie was one — "We strongly believe that PS Yandex does not take into account the data about which you wrote". Point.

Behavior in search results

We should also highlight the tasks performed:

the
1. "Employee" opens the search engine and sets the region (in my case, if he is not from Moscow).
the
2. Enters a search query.
the
3. Goes on a few random sites.
the
4. Goes to my website.

Very strange, when the user changes from Poltava region to Moscow and looking for concrete blocks last night when he was looking for GDZ for 8th grade. Next, the user clicks on two random website. Yeah. Most often it is the sites of the TOP 5 or TOP 10. Viewing the websites, the user quickly changes to a website with 29 seats and after the search finishes. Yes-a very natural behavior.

Light my mirror, tell me or watch Webvizor

"What is it?" — you will ask. And this is the scale of activity of the user of the Webvizor. She displays the typical behavior of the user from cheating on my website. According to my calculations so behaved 7 users out of 10. For those who do not know the meaning of these colors on the scale of an activity.



That is, the user passed to my site in 20 seconds “jerk” after you switched to another window and do something there 2 minutes. Then returned, somewhere clicked and closed the page. It is clear that most likely a worker performs several tasks simultaneously, and makes it to switch to another window.

And in the appendage to everything: in Webvizor, you can filter the visits of certain visitor that helped me to find another drawback of one user can run a job several times a week.

Result

After a difficult 4 months I decided to finish the Saga with this experiment. I won't bore with the story about the other problems, like incorrect removal of items, unclear execution of jobs and withdrawals, numerous small bugs in the interface (which they promised to fix, but this is not happening) and more.

It turned out that I spent over 4 months about 40 thousand rubles, a lot of time and nerve cells. In early may, a week after stopping the advance, I have recorded the following traffic:

the
1. A 220./ day.

Site B 180./day.

In an earlier post I promised not to compare positions, but, as a digestif should do it:



If you decided to use the cheat behavioral, aware of the shortcomings of "cheaters":

the
1. No ability to set depth and time of viewing.
the
2. When setting the desired region of promotion you will get people from all over Russia and the CIS.
the
3. Unnatural behaviour "employees" of the service on your website and on search results.
the
4. same "employee" can visit your website several times a week.

For myself, I decided that in the future use of questionable methods of grey will not. And you do not advise.
Thank you!

Article based on information from habrahabr.ru